Apart from the fact that I said "embryonic" twice, when I meant to say "symbiotic" (This will haunt me for a long time - I can only blame jetlag), I think it turned out rather well. The discussion, which can be accessed here, started around why there is a buzz around digital transformation now. I argued, as I usually do, that everything is getting more distributed, and we are getting onto new technology s-curves that can handle new data sets and digital business models. Poornima concurred, that tech, like machine learning, makes data that used to be out of reach, available. While this is the case, Dan Vesset added a sobering thought that only about a third of results of analytics are available for employees. He later also made the point that when he started as an analyst 18 years ago, the ratio of analysis to data prep was 20/80, and that's the same as it is today. So while technology has emerged furiously, our ability to consume it hasn't. No wonder then Qlik is so hung up on Data Literacy. Both Vesset and Poornima made the point that data ethnographers and anthropologists are starting to emerge to help interpret results from data scientists. I thought this was an interesting parallel to Valerie Logan's [i] points on data as different languages and dialects. And here's the thing. Human sciences have never been more important in our field, traditionally seen as a pretty clinical and logical.
And vice-versa. The discussion continued around how data-driven cultures need to be question-driven cultures. However, this needs to be constantly balanced with bias, which is deeply engrained into the human psyche. Poornima made an excellent point; "Let the data speak for itself, and lead you to the answers. Rather than cherry-picking data to suit your answers." Perhaps machine learning can help us here. Sure, there will be some automation, but, in Poornima's words - "That is in areas that shouldn't have been manual in the first place". Or in Dan Vesset's words; "There are a lot of different touch-points where machine learning can automate specific tasks within the analytics workflow". So, here's the paradox; AI will remove some of the bottle-necks, and finally help drive adoption and value, and as a result, will make data and analytics more human than ever. During my time as an industry analyst, I used to call it "reaching the other 75% by 2020", i.e. "information workers" and "workers with information". It will need to accelerate on the home stretch, but AI may very well save my prediction. So, what's next then? Well this was the final part of the panel discussion, but I'll leave that as a cliff-hanger for you to listen to the Virtual Panel.
[i] “Information as a Second Language: Enabling Data Literacy for Digital Society”